INDEX

Explanations

instances of being "lost" or "caught" in a variety of contexts

oai_token-act-pair · gpt-3.5-turbo

phrases that imply being "caught" or "lost" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

saf

-0.71

Phys

-0.64

available

-0.64

hey

-0.64

DragonMagazine

-0.63

flush

-0.63

nice

-0.63

kind

-0.62

 Fighters

-0.62

grad

-0.61

POSITIVE LOGITS

 limbo

1.15

 obscurity

1.06

 knots

1.01

mud

0.94

fog

0.91

 traffic

0.87

 bureaucracy

0.87

 confusion

0.86

 amber

0.85

 debt

0.84

Activations Density 0.155%

instances of being "lost" or "caught" in a variety of contexts

phrases that imply being "caught" or "lost" in various contexts

No Comments

No Known Activations